OpenAI Launches Bug Bounty Program for Abuse and Safety Risks

OpenAI Launches Bug Bounty Program for Abuse and Safety Risks

OpenAI has announced a public safety bug bounty program focused on AI-specific abuse and safety risks in its products, complementing its existing security bounty program. The Bugcrowd-run program accepts reports on prompt injection, data exfiltration, agentic product abuse, connector vulnerabilities, and exposure of proprietary information, with discretionary rewards up to $7,500 for high-severity, reproducible issues. #OpenAI #Bugcrowd

Keypoints

  • OpenAI launched a public safety bug bounty to address AI-specific abuse and safety risks.
  • The program complements the existing security bounty and is hosted on Bugcrowd.
  • In-scope issues include prompt injection, data exfiltration, agentic product misuse, and connector weaknesses.
  • Researchers are encouraged to report flaws in Atlas Browser, Codex, Operator, Connectors, and other ChatGPT tools.
  • Rewards up to $7,500 may be granted for high-severity, reproducible issues with clear remediation steps.

Read More: https://www.securityweek.com/openai-launches-bug-bounty-program-for-abuse-and-safety-risks/